ghc 9.2 wip #43
ghc 9.2 wip #43
Conversation
|
in case it helps, http://blog.shaynefletcher.org/2021/05/annotations-in-ghc.html gives a summary overview of annotations in 9.2.1 also, depending on your needs, you might get by with the native record dot syntax support offered in that release |
|
I've checked the native record dot syntax, but it's really incomplete, and seemingly won't make it into 9.2 in such an useful form that record-dot-preprocessor is. |
|
Thanks for the code! A very useful start. Any information I could provide that would let you get any further? Or would you prefer to just stop there, and hopefully this can serve as a reference when someone else gives it a go? |
|
GHC 9.2 is now out, if @l29ah or someone else wants to pick this up. |
Depends on - https://ghc.gitlab.haskell.org/head.hackage/ - well-typed/generics-sop#147 - well-typed/json-sop#6 Unfortunately, RDP does not work (WIP PR at ndmitchell/record-dot-preprocessor#43) so it must be disabled for the `large-records` test suite, which depends on - #38 Build with ``` cabal build --allow-newer=base --allow-newer=template-haskell --allow-newer=time --allow-newer=ghc-prim ```
Depends on - https://ghc.gitlab.haskell.org/head.hackage/ - well-typed/generics-sop#147 - well-typed/json-sop#6 Unfortunately, RDP does not work (WIP PR at ndmitchell/record-dot-preprocessor#43) so it must be disabled for the `large-records` test suite, which depends on - #38 Build with ``` cabal build --allow-newer=base --allow-newer=template-haskell --allow-newer=time --allow-newer=ghc-prim ```
Depends on - https://ghc.gitlab.haskell.org/head.hackage/ - well-typed/generics-sop#147 - well-typed/json-sop#6 Unfortunately, RDP does not work (WIP PR at ndmitchell/record-dot-preprocessor#43) so it must be disabled for the `large-records` test suite, which depends on - #38 Build with ``` cabal build --allow-newer=base --allow-newer=template-haskell --allow-newer=time --allow-newer=ghc-prim ```
|
#50 cleaned this up and finished it off, thanks all! |
re #42
i've mauled some terms in Compat to typecheck, using the relevant ghc commits as a reference, without much understanding of the semantics involved, but then i'm faced with a much bigger volume of type errors in the preprocessor itself, so i'm not sure i will complete it